A Mixed-Integer Linear Programming Model for Minimizing Production Costs with Soft Constraints: A Case Study of the Fast-Moving Consumer Goods Industry

บทคัดย่อ

High midstream supply chain costs related to production, by which financial liquidity is directly affected, are addressed in this research. An optimization model for monthly aggregate production planning was developed to reduce costs in the fast-moving consumer goods industry using linear programming (LP). Initially, the model (Minimize Z1) was presented to the factory and at a conference. Knowledge was integrated so that the efficiency of the production plan could be improved for the lowest cost. Based on industry feedback, the study was expanded to daily production scheduling. Outputs from Minimize Z1 were used as constraints in the daily model, and variables for the on-off status of production lines were added so that actual working conditions could be matched. It was shown by the evaluation results that traditional factory methods were outperformed by the developed models in every year. Average inventory costs were reduced by 99.98%, average raw material costs by 8.46%, and total average costs by 85.8 million baht (approximately 2.6 million USD) by the LP model (Minimize Z1). Furthermore, regular and overtime working hours were efficiently allocated by applying mixed-integer linear programming (MILP) in the second model (Minimize Z2). As a result, average daily costs were decreased, and the management of operational resources was emphasized. When regular and overtime costs were compared, it was found that costs were further reduced by Minimize Z2 by a total of 8,270,382 baht (about 255,000 USD) from the previous results. An important contribution to knowledge was made by this finding. Also, a p-value of less than 0.05 was found through a statistical test using 72 months of data. It was shown by this test that costs were significantly reduced by the Minimize Z2 model.